Fault Detection and Diagnosis in Stirling Engines: A Computational Approach

Fault detection and diagnosis (FDD) are critical for ensuring a reliable operation of Stirling engines, which are known for having an increased interest in research in various applications ranging from power generation to marine engines. The current study investigates advanced computational approaches for FDD in Stirling engines, leveraging machine learning and signal processing techniques. By analyzing sensor data and system parameters, these methods aim to detect anomalies indicative of faults and facilitate timely maintenance interventions to prevent operational disruptions and costly failures. The proposed methodology involves data collection from an accelerometer that monitors vibration signals, followed by feature extraction and applying a Blind Source Separation (BSS) using a Fast Independent Component Analysis (Fast ICA) algorithm. A case study demonstrates the effectiveness of the computational approach in detecting and diagnosing faults in Stirling engine systems. The discussion explores the challenges and opportunities associated with implementing automated FDD in real-world Stirling engine applications with the use of BSS. Overall, this research contributes to advancing the reliability and performance of Stirling engines through enhanced fault detection and diagnosis capabilities, offering insights for both researchers and practitioners in the field.